Adding a Dryrun

If you are contributing a system or experiment to our code repository you must add a passing dryrun test to the .github/workflows/run.yml file before your pull request will be merged.

Recommended systems/experiments to use as tests:

• genericx86

• tioga

• saxpy

• amg2023

For example, if you are contributing a system called foo you would test it with existing Saxpy experiment, the hash ID will be generated for you during setup and output by benchpark:

- name: Dry run dynamic saxpy on dynamic foo
  run: |
    benchpark system init --dest=foo-system foo
    benchpark experiment init --dest=saxpy-openmp saxpy +openmp
    benchpark setup ./saxpy ./foo-system workspace/
    . workspace/setup.sh
    ramble \
      --workspace-dir workspace/saxpy/foo-{hashID}/workspace \
      --disable-progress-bar \
      --disable-logger \
      workspace setup --dry-run

If you are contributing a benchmark and/or experiments to our code repository you can use an existing system to test your benchmark and experiments.

For example, if you are contributing a new benchmark called bar:

- name: Dry run dynamic bar on dynamic genericx86
  run: |
    benchpark system init --dest=x86-system genericx86
    benchpark experiment init --dest=bar-benchmark bar
    benchpark setup ./bar-benchmark ./x86-system workspace/
    . workspace/setup.sh
    ramble \
      --workspace-dir workspace/new-benchmark/genericx86-{hashID}/workspace \
      --disable-progress-bar \
      --disable-logger \
      workspace setup --dry-run